cancel
Showing results for 
Search instead for 
Did you mean: 

General Discussions

mackbolan777
Forerunner

Can I use the displayport for video and hdmi for audio via my A/V?

I'm having random crash issues in game even before I did this. I have a new monitor that needs to use the displayport for 185hz, so to keep my audio, I used an hdmi port to my A/V receiver. Is this a problem or still poor driver issues? Running the 20.8.3 driver now, most stable for me but still crash problems.

The error is "processor failure" and "processor hierarchy error" with some ACPI number, always different. The PC will either crash to desktop, reboot, or green screen to reboot.

Hardware wise all checks out in DXDiag, PSU is 1000w up to spec, CPU is new Ryzen 5 3600X, AsRock X570 Phantom Gaming 4S, 16GB TeamForce Gaming DDR4 3733 RAM, Seagate Compute 1TB SSD, Corsair H110i GTX V2 cooler, Gigabyte Windforce  RX 5600 XT Gaming OC 6GB(14Gbps BIOS). This audio thing is the last thing I can think of causing a possible issue. Happens after playing BF4 for an hour or longer, pretty much any game that long will cause it. The 20.9.2 driver made it worse. Losing faith in AMD over this.

"It worked before you broke it!"
0 Likes
1 Solution

I have a similar setup with my Onkyo A/V connected via HDMI to my GPU card & my 4K monitor connected via DP . 

Both works flawlessly without problems. I use my Onkyo A/V for 5.1 Audio. 

This combo worked with both my previous AMD Gpu card and my present Nvidia GPU card.

Your problem most likely is either a driver or Gpu or configuration or hardware issues.

Have you checked to see if anything is overheating?

Download OCCT and run all 3 stress tests.

See if it crashes during the CP U test.

Keep a close eye on temperature and PSU outputs.

Same goes for the GPU & PSU tests

Check Device manager. For any errors and Windows event viewer under errors.

In DXDIAG check the last category Diagnostic to see what files are having issues

 

 

 

 

 

View solution in original post

13 Replies

I have a similar setup with my Onkyo A/V connected via HDMI to my GPU card & my 4K monitor connected via DP . 

Both works flawlessly without problems. I use my Onkyo A/V for 5.1 Audio. 

This combo worked with both my previous AMD Gpu card and my present Nvidia GPU card.

Your problem most likely is either a driver or Gpu or configuration or hardware issues.

Have you checked to see if anything is overheating?

Download OCCT and run all 3 stress tests.

See if it crashes during the CP U test.

Keep a close eye on temperature and PSU outputs.

Same goes for the GPU & PSU tests

Check Device manager. For any errors and Windows event viewer under errors.

In DXDIAG check the last category Diagnostic to see what files are having issues

 

 

 

 

 

Good to know someone else has been using a separate port off the GPU for audio via A/V and no related issue. I've run Prime 95 for 30 minutes, but don't like the temps the CPU hits. I know that test is primarily CPU but it does stress the PSU and system overall. I have custom RAM settings so I have run the paid version of Memtest 64 overnight several times to make sure all is well. I've tried Superposition and no failure. I'll have to download OCCT again and try it but this problem is so weird, it seems gaming exclusive. Now, last night I played BF4 on Ultra 960mv/1820 core clock/1800 VRAM custom fan curve all via Adrenaline 20.8.3 and had no issues for the 2 hours of play. I then watched a 1.5 hr. movie. The movie doesn't really stress anything. 

I'm still thinking it could be heat related. Even though the temp of VRAM 70c, Tj 70c, under full load is "ok" in AMD's eyes, it is sustained 158F. To me, it just seems high, others say "wow I wish my card ran that cool". Those temps boil down to cooler design. I watched a tear down of my card and a the 5700 XT version on Gamers Nexus and the cooler is really junk. The heat pipes don't fully engage the GPU die and the pads on the VRAM are less than desired. That said, I've seen liquid cooled 5700 XT's with little to no change in operating temp from air! One thing I may try is removing the shroud and the stock fans that rattle at 2300 RPM or more and replace those with a pair of high flow PWM's and no shroud obviously. I saw a reddit post and he claimed temps were more around 50-60c under load across the entire board. He used 2 x 140mm Noctua's and 2 X 50mm Noctua's over the VRAM on a 5700 XT. I just need 2 since I have the 2 fan version of the RX 5600 XT, which has a plastic backplate not metal like most others.

I'll skip repasting the card because I saw that Thermal Grizzly and new pads didn't make a difference and pad thickness is paramount to get right. It's just almost cheaper to buy a new card at that point. The error codes are for "processor core failure" and "processor Hierarchy error", I read these are actually triggered by the GPU but I've seen posts about Intel/Nvidia guys having a similar issue. The common part is all of us running Windows 10 x64 version 1903-2004. Enter a Windows bug, and that's a whole other rats nest. Thanks for replying, I'll let you know if OCCT turns up anything or if I ever find a solution other than buying a 2070, lol.

"It worked before you broke it!"
0 Likes

Memtest is excellent for detecting defective RAM but not if RAM is incompatible or not.

As for temps on the rx5700 GPU cards and possibly all of the rx5xxx GPU cards the Maximum Operating Temperature is 110C.

The rx5700 has several thermal sensors throughout the GPU card. If any of the thermal sensors reach 110C it will start to throttle to lower the temperature below 110C.

I suggest you do a CMOS clear to reset your BIOS back to default to see if your problem is a BIOS setting.

Also run and stress tests using OCCT.

Been there, done that. I have the paid version of Memtest due to a prior issue with running my ddr4 3733 on this board: AsRock X570 Phantom Gaming 4S, which since a BIOS update, DRAM CALC, and Memtest overnight (1000% coverage), the RAM is ok. Just ran OCCT for an hour and 1 WHEA Processor Core, unknown source, APIC ID:0, corrected. Then looking further there was a few more OCCT didn't show:GPU3DDX11.exe application failure for OCCT when I tried running the 3d test alone. The 2nd try it worked no error. This problem is weird for sure. All voltages are perfect, no crazy OC, tried default everything before and same issue just longer to reproduce.

"It worked before you broke it!"

BTW, mine is the RX 5600 XT, not that it matters for point of temps. I know what AMD rates as temps, I still think 70c is a bit crazy hot (158F) for the VRAM part. Hotspot runs at 80c. So AMD would say thermals are fine. Before I adjusted the fan curve and got rid of "zero RPM" and took the voltage down to 960mv max, the card would hit 90c, VRAM 90c, Hotspot 100c and crashing was much more frequent.

Last night I ran DDU and installed 20.8.3, then played BF4 for about 2 hours and watched a movie with no problems. Looking back in Event Viewer, there were some WHEA's relating to processor core/bus interconnect error but corrected. I can't get the CPU to fail a test to the point of crash though. When it crashes, there's WHEA's for processor core, processor hierarchy, sometimes AMD Radeon Software.exe, all in the same timeframe.

For 20+ years I've built PC's and the only ones with weird issues were AMD builds. I had a choice last year to go Intel/Nvidia or try Ryzen. Seems like I made a mistake, even though this build outperforms my previous i7 8700K/Nvidia 1080i build. I'd still be running that had it not got stolen. Intel was easy, click n go, done. This is like fine one day, a disaster the next with no real trail of what's causing the issues. These same complaints run across this forum, Reddit, the entire net. Why no one can figure out what is truly causing it is just nuts. In fact I can't find one AMD user that can honestly claim they never had this or a similar issue with Ryzen 5 builds and an AMD anything card. Intel builders with AMD cards seem to experience similar crashes, so again leads to the cards or drivers being an eternal issue.

The one thing that greatly reduced the crashes was installing only the driver and using 3rd party software to set fan curve, clocks, etc. It took the issue down to once a week or two from the total randomness it is now. I can go back to that way but figured I'd try Adrenaline one more time to see if it's the software itself and not the driver per se. When I see Processor Core/ Bus-Interconnect Failure, that makes me question the platform as a whole. Going back to the AM3+ CPU's like the Bulldozer or the 8350, there weren't as many issues. Some video drivers but not like this. 

I'd be willing to go to the 5700 XT if that was the cure but they have the same issues. I'd try the 3600XT or better yet the 5600X but if I still get these errors, man I would lose it. I'm waiting for those who spent $650 or more for a 6800XT to have this happen, then I'll know where to put my future money. At least I know hooking my A/V to an HDMI port for audio is not the problem. One thing out of the mix. PSU is clear too. It gets really frustrating when parts are super expensive and in 2020/21 we're locked into "hip shot" diagnostics of trial and error. 

I know a lot of people will just say "yep, I reset the BIOS", just to get past that rhetorical question, however, I do very thorough diagnostics since I ran my own business for 20+ years and gamed all the way to sponsorship at one point. So you can trust me that I've done everything in the book and out of the box to figure this out, except build a new Intel/Nvidia PC to see if it happens. Or spend a lot of money on an Nvidia card to see if it might be this platform as a whole. I could even try Linux but there's very few games to test with, let alone drivers, but that would rule out the OS. 

Thanks for responding/reading, sometimes it helps to "get it all out". Whew.

"It worked before you broke it!"

Open a AMD Service Request ticket (Official AMD SUPPORT) and see what they recommend you do from here: https://www.amd.com/en/support/contact-email-form

Maybe they might suggest to RMA your GPU card or CPU. just hypothesizing. But if you removed the Cowl it is possible your Warranty is voided by the manufacturer.

Note: I tend to suggest the basic troubleshooting suggestions since the most experience troubleshooter can overlook the simplest troubleshooting step.:smileywink:

I would suggest the same steps as well. Yes, it's easy to overlook one test. Now, the thing is making me a liar. 2 days and no crash, even with the one WHEA error while running OCCT, Processor Core/Bus Interconnect/corrected. Played BF4 on Ultra for 2 hours last night, no issues and thermals showed the VRAM junction riding 70c the entire time. 160 adv. FPS

One odd thing is the current draw. Running OCCT, it shows the card pulling 107-111A at times, my PSU says 83.5A on the 12v rail. How is it that it goes over the spec? Wattage was peaking at 160w, normal for load. CPU power package 105w peak, 93w steady load. 90A though. Corsair must really under rate their PSU's. RM1000w

"It worked before you broke it!"
0 Likes

Try using a different Monitoring software to see if you get the same results. Might just be OCCT monitoring software which I believe is based on HWMonitor monitoring program.

What does Radeon Wattman show compare to what OCCT shows?

Wattman is the GPU program of Radeon Settings.

Also you can check your CPU and Motherboard by using AMD Ryzen Master.

0 Likes

OCCT and HWINFO show Amps, Watts, and Volts. Watts are ok, voltage is in spec, Amps are crazy high but everything I Google on it says little about Amps period. 12v rail, 83.5A, I didn't look at how many Amps are on the 3.3 or 5v part, so maybe that Amp reading is a combined thing and certainly not all at once. The card using 160W is totally normal, CPU using 105w peak and 90A is also ok. PBO spikes the wattage but these 3600X's are 2 cores and 1 CCX light of a 3900X. CPU calms down to 63-70W steady load. Ryzen Master would never load for me. Says I need Windows 10, I have Windows 10 x64 Home. Tried it with Pro, same deal. Tried it with 1903 and 2004, same deal. I gave up on it.

"It worked before you broke it!"
0 Likes

To tell you the truth I am out of suggestions.

I would contact AMD Support and give them your data and see what they say.

The high amperage draw could be a glitch in the monitoring software or a Sensor not working correctly.

I sent them a DXDiag before with the crashing back in June and after 2 weeks they get back to me with nada. It's an odd thing. Now going on 3 days and it seems fine. No crash yet, just the one error in OCCT. May be this driver is the best one for me. Running 20.8.3 with the Adrenaline software. If it acts up again, i can try running the base driver only or get a sound card to run an optical to my A/V. I did a "print screen" for another guy on here and it shows as 2 desktops. My back round printed next to my actual desktop as if I had 2 monitors. The right side "monitor" showed just my back round, no icons. The left side was my active desktop. So running the HDMI to my A/V actually sends a video signal that the PC sees but I don't. I can assume that when I game, the extended desktop is just a static back round and not rendering the game but it's still extra power/resources from the card one would think. Here's a screen grab. I only use "Hudsight" for BF Vietnam because the card won't render a gun sight in that 2004 game,lol.desktop.png

"It worked before you broke it!"

My A/V isn't connected to any monitor just the 5.1 speakers even though in Windows it shows as being connected to a monitor.

I had something similar when i connected both HDMI and DP to my monitor at the same time I had the second picture to the left you are showing. Except I couldn't go into my desktop and no programs would open up even though in task bar it showed it was opened.

I couldn't understand why my Desktop wasn't showing any icons but in the task bar apps were opening. Then I went to Windows Settings - Display and notice I had two monitors connected and realize that my DP was replacing my HDMI connection which had the active Desktop.

Once I disconnected the DP cable from my monitor, I got my HDMI active Desktop back again. I felt really foolish since I went as far as doing System Back recovery which of course didn't fix the issue.

Well even more interesting is I ran 3D Mark Timespy without the HDMI and then with the HDMI, no audio either time and without the HDMI I picked up 4 FPS 7924 vs. 7920. I ran BF4 using headset and my game was smoother, FPS more steady at 200 high 130 low 157 avg on Ultra at 1080p. Temps were the real interesting part in that the both GPU and VRAM hit 70c but cooled to 56c during less intense times. With the HDMI in the temp stays at 70-78c on the VRAM and the GPU is steady at 66-70c, neither cooling off during less intense parts of the game. So I would say that running audio off an empty port might be ok but one is losing 4 FPS or more in my case and the card is using more power to the point of staying hot longer. 

Could having a static desktop with a video signal going to a blackhole cause corruption at some point causing the crash? If Windows sees the A/V as a "display" and it's sending a 1080p full screen image to an HDMI port with no way out perhaps the PC gets confused intermittently? No one has really delved into this angle. It's not a typical multi monitor config.  I don't think AMD nor Nvidia thought about anyone doing this setup.

When I built this I ran this card via HDMI to the A/V passthrough to my HDTV. I then realized at some point, that at 120hz max refresh, that was my max FPS, I thought. But it got worse, the A/V only allows for a 60hz passthrough meaning I was only getting 60 FPS showing no matter what any FPS counter said. Connecting it directly to the HDTV, my FPS was better but the 5ms GTG was not great with input lag. So I recently bought this Sceptre 32" 1800R 1080p with a 1ms GTG, FreeSync, and a 185hz via DisplayPort. The video passthrough on the A/V won't work with this monitor even in HDMI config. Night and day difference off DP, except color reproduction isn't hot on these type of monitors for movies and such. An IPS would be for that but than you give up refresh rate for looks.

So now I'm thinking a sound card because this expensive AsRock X570 didn't come with an optical out, just crappy Realtek 3.5mm jacks which won't do 5.1, let alone the 7.1 they claim. I should've went with the Logitech Z-906, they no longer make but on Ebay you can find them. But I bought a Yamaha RXV-371 because I wanted the HDMI passthrough not realizing it had a refresh rate limit of 60Hz. I hooked that to my Insignia 100W bookshelves and bought a Rockville 8" sub. For the money, the Logitech's make more sense now. All this to get 192kz that you never truly get out of PCM, as if my 47 year old ears could even tell the difference between 192kz and 88kz, 48, yes. 16 bit vs. 24bit yes, but the rest no. Heck my Corsair Void's sound fantastic but my neck gets sore from the weight after an hour. My vertebrae's at C2-3 are almost fused, longer story... 

The point is the Logitech's would've connected to the 3 jacks on the mobo and I would've had 5.1 in THX, woohoo! Nope, went the hard way. Tried to save a few bucks, either way it cost the same and I have a louder 3.1 that fakes 5.1. I guess $60 to find out if it ends the mystery is worth it. Like I said, this monitor won't work via HDMI passthrough on my A/V, it displays at 640 x 480. So might as well use the DP and get a sound card with optical to connect the A/V. The headphone jack works but the sound quality is horrible, Realtek's software is junk and it has no depth. The HDMI sound is great. Headphone jack means turning the volume up on the A/V to -30 barely any bass on HDMI -50 is decent, -40 you feel in your chest, -19 and the neighborhood hears it.

If I knew all I was losing is 4 FPS, I would just use the HDMI but my tests show it's more than that. Maybe AMD can make a feature to run audio via HDMI and shutdown the video part when using the DP or similar setups. It does make sense that video would be sent too. A good question would be how many have a crash issue and the sound is hooked up the same way. It seems stupid but a display is a separate process and it's tossing audio over both outputs too. The monitor being speakers off. VRAM, GPU, RAM and CPU all using some resources to a blackhole for audio.

"It worked before you broke it!"